Report Compiled: 2020-06-15

Johns Hopkins Repo Commit: 47885a1 Max Data Date: 2020-06-14

NYT Repo Commit: e6dc96a Max Data Date: 2020-06-14

This is an automatically generated report containing analyses of the COVID-19 epidemic in Iowa and elsewhere. All models under consideration here are stochastic SEIR compartmental models, fit using Approximate Bayesian Computation using the ABSEIR software for R. Source code available upon request. Questions should be directed to grant-brown@uiowa.edu

There are two general classes of model:

  1. Models which use a single location of mortality data to estimate the epidemic curves
  2. Models which are used to provide informative prior information about the epidemic in Iowa, based on analyses of other locations.

In both cases, we have to make assumptions about the shape of the underlying contact distributions. Namely, can we assume that contact in each location shifted within one week of governmental action, or is the shape of the curve more complex. With that in mind, we look at both types of models. In addition, for the State of Iowa, we consider whether or not it is most reasonable to assume that intervention efforts began on 3-17-2020 or 4-4-2020, which correspond to the emergency declaration and the official closing of schools.

** A huge array of models are presented in this document, and they are not reviewed by our team of experts before posting to this page. Some of these are guaranteed to be inadequate or misleading if interpreted by themselves. These results should be considered raw material for follow-up reporting, investigation, and decision-making.**

R0 Summaries: Single Location Analyses

Mortality Estimates: Single Location Analyses

Here, we present the compared results of analyses of the COVID-19 outbreak in a number of locations. We begin by comparing the estimated posterior distribution of mortality rates in each location.

Model Fit: Single Location Analyses

In the following tabbed sections, we present diverse output from the single location models, including projections over time of the following important quantities:

Illinois

Washington

Minnesota

Iowa (3-17)

Iowa (4-4)

Iowa (workplace mobility)

Iowa (retail mobility)

Illinois (Spline Model)

Washington (Spline Model)

Minnesota (Spline Model)

Iowa (3-17, Spline Model)

Iowa (4-4, Spline Model)